Development and validation of a prognostic model for 90-day survival in patients with alcohol-associated cirrhosis and acute decompensation.

Abstract

Patients with alcohol-associated cirrhosis and acute decompensation are considered critically ill and have a higher risk of short-term mortality. This study aimed to establish a nomogram to evaluate their 90-day survival and identify factors that affect disease progression. We included patients from September 2008 to December 2016 (n=387 in the derivation group) and from January 2017 to August 2020 (n=157 in the validation group). LASSO regression and Cox multivariate risk regression were used to analyze the influencing factors of the 90-day mortality risk, and a nomogram was constructed. The performance of a model was analyzed based on the C-index, area under the receiver operating curve, calibration curve, and decision curve analysis. Total bilirubin >10 upper limit of normal, high-density lipoprotein cholesterol, lymphocyte and monocyte ratios ≤2.33, white blood cells, and hemoglobin were identified as independent risk factors affecting the 90-day mortality risk of patients and the nomogram was developed. A nomogram demonstrated excellent model predictive accuracy in both the derivation and validation cohorts (C-index: 0.976 and 0.945), which was better than other commonly used liver scoring models (p<0.05). The nomogram also performed good calibration ability and more clinical net benefit. According to the nomogram score, patients were divided into high- and low-risk groups. Mortality was significantly higher in the high-risk group than in the low-risk group (p<0.0001). The nomogram could accurately predict the 90-day mortality risk in patients with alcohol-associated cirrhosis and acute decompensation, helping to identify high-risk patients and personalize treatment at their first admission.